一种多边协商的群智感知网络协同防御模型  

Multilateral Negotiation Cooperative Defense Model for Crowd Sensing Networks

作　　者：崔广金 赵国生[1] 王健[2] CUI Guangjin;ZHAO Guosheng;WANG Jian(School of Computer Science and Information Engineering,Harbin Normal University,Harbin 150025,China;School of Computer Science and Technology,Harbin University of Science and Technology,Harbin 150080,China)

机构地区：[1]哈尔滨师范大学计算机科学与信息工程学院,哈尔滨150025 [2]哈尔滨理工大学计算机科学与技术学院,哈尔滨150080

出　　处：《小型微型计算机系统》2024年第8期2008-2016,共9页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61202458、61403109)资助;黑龙江省自然科学基金项目(LH2020F034)资助;哈尔滨市科技创新研究基金项目(2016RAQXJ036)资助.

摘　　要：群智感知网络中的攻击容易造成大面积感知任务中断,降低群智感知网络的可用性.针对群智感知网络中孤立的感知节点由于自身资源不足无法抵御攻击导致网络系统崩溃的问题,提出一种基于多边协商的感知联盟生成方法协同完成防御任务,提高感知网络的防御能力.首先,对感知节点群体行为和关系进行形式化建模分析,将感知节点泛化为三类不同角色,对应完成不同的防御任务.然后将感知节点对防御任务的探测优势、距离优势和能量优势转化为感知节点对攻击的防御优势.最后,提出以对手建模为中心的基于多边协商的感知联盟协作策略,各感知节点通过多边协商参与防御联盟的生成,而获得联盟收益激励.仿真结果表明,随着让步权重值的增加协商成功率逐渐增大,但是感知节点防御收益却逐渐减小达到均衡.与随机的联盟生成方式在生成协同防御联盟的纳什距离和防御联盟效用方面进行比较,验证了该方法完成防御任务的高效性.Attacks in swarm intelligence awareness networks are prone to interrupt a large area of awareness tasks and reduce the availability of swarm intelligence awareness networks.Aiming at the problem that isolated sensing nodes in swarm intelligence sensing network cannot resist attacks due to insufficient resources,a multi-lateral negotiation based perception alliance generation method is proposed to cooperate to complete defense tasks and improve the defense capability of sensing network.Firstly,the group behavior and relationship of sensing nodes are formally modeled and analyzed,and the sensing nodes are generalized into three different roles,which correspond to different defense tasks.Then the detection advantage distance advantage and energy advantage of the sensing node against the defense task are converted into the defense advantage of the sensing node against the attack.Finally,a collaborative strategy of perception alliance based on multilateral negotiation with adversary modeling as the center is proposed.Each sensing nodes participates in the formation of defense alliance through multilateral negotiation and gets the incentive of alliance income.The simulation results show that the negotiation success rate increases gradually with the increase of the concession weight value,but the defense benefit of sensing nodes decreases gradually to reach equilibrium.Compared with the random alliance generation method in the aspects of Nash distance and effectiveness of cooperative defense alliance generation,this method is proved to be efficient in completing defense tasks.

关 键 词：群智感知网络 协同防御 对手建模 多边协商 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]